Edmund Zagorin joins Kevin Appleby on the GrowCFO Show. Edmund is CSO of Arkestro. Arkestro helps your team make buying decisions, faster, and at scale using Predictive Procurement Orchestration



Kevin and Edmund explore anchor bias and consider how CFOs can help their procurement teams avoid anchor bias and use predictive modelling to make better buying decisions.



Anchor Bias and Predictive Modelling



Anchor bias, also known as anchoring bias, is a cognitive bias that refers to the tendency for individuals to rely heavily on an initial piece of information (the "anchor") when making decisions. In the context of procurement, anchor bias can significantly impact the effectiveness of predictive modeling.



Predictive modeling in procurement uses historical data and statistical algorithms to forecast future outcomes. However, if the initial data or 'anchor' used in the model is biased or inaccurate, it can skew all subsequent predictions, leading to flawed decision-making and potentially costly mistakes.



There are two main types of anchor bias: self-generated and externally provided. Self-generated anchor bias occurs when the individual creates the anchor based on their own knowledge or assumptions. Externally provided anchor bias, on the other hand, is when the anchor is provided by an outside source, such as a vendor's initial price quote.



For example, if a procurement professional uses the cost of a previous contract as an anchor when forecasting future costs, they might ignore changes in market conditions, material costs, or supplier capabilities that could lead to higher or lower costs. This could result in budget overruns or missed opportunities for savings.



Similarly, if a supplier's initial price quote is used as an anchor, it could influence the procurement professional's perception of what is a reasonable price, potentially leading to overpayment.



How to Mitigate Anchor Bias in Predictive Modelling



To mitigate anchor bias in predictive modeling, procurement professionals should consider the following strategies:




Use Multiple Data Points: Instead of relying on a single piece of information, use multiple data points to create a more accurate prediction. This could include data from different suppliers, contracts, or time periods.



Challenge Assumptions: Regularly question and validate the assumptions that underlie your predictive models. This can help identify any potential biases and correct them before they impact your forecasts.



Seek Diverse Opinions: Consult with colleagues or industry experts to get different perspectives. They might provide additional insights that can help adjust your anchor.



Train and Educate: Provide training and education on cognitive biases for procurement staff. Understanding these biases can help individuals recognize and mitigate them in their own decision-making processes.




While anchor bias can pose a significant challenge in predictive modeling for procurement, it can be mitigated through awareness, careful data analysis, and ongoing validation of assumptions. By doing so, procurement professionals can make more accurate predictions and better decisions.



Edmund shared his procurement, strategic sourcing, and data science background, and discussed the importance of collaboration with business stakeholders to achieve the best outcomes for all parties involved.



Edmund and Kevin discussed anchor bias in procurement negotiations and how the use of predictive models can help avoid it. They also talked about the impact of first offers on price variance and the asymmetry in technology and data between buyers and sellers.



Edmund and Kevin discussed how predictive models powered by AI can enable procurement and finance teams to negotiate with suppliers more effectively by using behavioral analysis instead of traditional cost or value frameworks. They also talked about the challenges in the relationship between procurement and finance,